According to Driver (1996):
‘The cluttered scenes of everyday life present more objects then we can respond to simultaneously, and more often we can perceive full at any one time. Accordingly, mechanisms of attention are required to select objects of interest for further processing. In the case of vision, one such mechanism is provided by eye movements, which allow us to fixate particular regions so that they benefit from the greater acuity of the fovea’.
Posner (1978) found that when people are told to fixate on one part of the visual field, it’s still possible to attend to stimuli seven or so degrees either side of the fixation point. Also, attention can be shifted more quickly when a stimulus is presented in an ‘expected’ rather than ‘unexpected’ location.
Thus visual attention isn’t confined to the part of the visual field which is processed by the fovea, but can be shifted without corresponding changes in eye movements. Indeed, shifts in attention precede the corresponding eye movement (Anderson, 1995). Posner (1980) calls this covert attention (also known as the spotlight theory).
Evidence in support of spotlight theory came from a study by LaBerge (1983). He presented participants with five-letter words and a probe requiring them to respond as quickly as possible. The probe (or target) could appear in the position of any of the five letters in a word.
In one condition participants were asked top respond to the whole word by categorising it. In another condition they were asked to respond to the middle letter only by categorising it. It was expected that, when being asked to respond to the middle letter, a narrow attentional beam would be employed.
The results of the study indicate that where participants were asked to respond to the word and their attentional beam was broad, it did not matter at which point on the display the visual probe occurred, response times were equally fast. However, when participants focused on the middle letter, the position of the probe was critical. Response times were significantly slower where the probe had not been presented in the centre of the five-letter display.
LaBerge demonstrated the probe would be responded to more quickly when if fell within the attentional spotlight than when it did not. He also demonstrated the attentional spotlight can have a narrow beam (letter task) or a broad beam (word task).
According to spotlight theory of visual attention there should be minimal processing of visual stimuli falling outside the spotlight.
Studies have confirmed this predicted. For example, Johnston and Dark (1986) investigated the time taken by participant to identify words which were first out of focus, but gradually became clearer. They discovered that presenting the same word or word of similar meaning to an unattended part of the visual field, immediately prior to the test word, usually had no positive effect in identifying the test word. This suggests, in line with the spotlight theory, no semantic processing of the unattended words.
However, other studies have shown the spotlight theory of attention to be over simplified. Juola et al. (1991) conducted a study involving a target letter being presented in one of three concentric circles#: the inner, the middle and the outer ring.
Participants fixated on the circles and were given a clue which gave them information about which ring the latter letter would appear in. If the zoom spotlight theory is correct. Speed and accuracy would be greatest for targets presented in the inner circle rather than the outer rings. Performance was best when the target was presented in the ring that had been cued. This does not support the spotlight theory which suggests a strong centre spotlight.
In a standard visual search experiment, the participant is looking for one target item in a display containing some number of distracting items. One of the attractions of this paradigm is that it brings a very common real-world visual behavior into the lab.
In our day-to-day life, we frequently look for the keys, the socks, the can opener, or any of a host of visual targets in the cluttered visual display that is our world. For example, a face in the crowd, and name in the phone book.
The efficiency of a visual search can be assessed by looking at changes in performance; generally reaction time (RT) or accuracy, as a function of changes in the "set size"; the number of items in the display. These changes, in turn, can be used to make inferences (i.e. conclusions) about vision without attention.
This simple paradigm allows researchers to examine how visual stimuli are differentiated, what stimulus properties attract attention, how attention is deployed from one object to the next, & how one keeps track of what was attended.
Not surprisingly, the visual search paradigm has been used extensively. Laboratory versions typically use highly artificial stimuli (coloured line segments, letters, etc).
In a typical lab study, participants would perform many searches for such targets amongst a variable number of distracters. The total number of items in the display is known as the set size. The target is presented on some percentage of the trials, typically 50%. Participants press one button if the target is present and another button if only distracters appear.
Searching for targets in a visual array is an important everyday task. Experiments have found that visual search involves two processes, the first being fast and efficient and the second being slower and less efficient.
However, studies of visual search have only been conducted in a laboratory environment, using relatively simple stimuli and we must therefore be cautious in drawing conclusions from such research for visual search in our everyday lives.